The measurement everybody said was missing exists: in SEM cells KMT2A::AFF1 occupies 20% of unmethylated CpG sites, CXXC1 50% and KDM2B 89%, and only the fusion tracks menin (r2 = 0.96)
also called Kerry 2017 re-read · MLL-AF4 versus CFP1 versus KDM2B occupancy in SEM · uCpG is not the limiting resource
A 2017 paper about a different drug quietly measured, in human infant-leukaemia cells, where the KMT2A::AFF1 fusion sits on DNA and where two rival proteins that read the same DNA signal sit. The fusion uses only a fifth of the available sites, its rivals use half and nearly all of them, and what marks out the fusion's fifth is the presence of menin.
The same thing, for a scientist
**This entry corrects a counted absence written one cycle ago in this graph and reverses part of its verdict.** `absence-no-experiment-tests-cxxc-reader-competition-at-a-kmt2a-fusion-target` asserts *"of the eleven human CXXC-domain proteins other than KMT2A itself, 0 have a published measurement of their chromatin occupancy in the same cells as a KMT2A fusion protein; 0/11, rule of three <27.3%"*. **At least two do, and the paper is already cited 29 times in this graph.** Kerry et al., *Cell Rep* 2017, performed Bio-CAP-seq (biotinylated CXXC affinity purification, a direct genome-wide readout of non-methylated CpG), ATAC-seq, and ChIP-seq for MLL(N), AF4(C), menin, PAF1, LEO1, **CFP1/CXXC1 and KDM2B — all in SEM, a human t(4;11) KMT2A::AFF1 B-ALL line, i.e. the index fusion in the index lineage** [E1]. The numbers are in the running text of Figure 1 and Figure S1F and appear nowhere in the title or abstract, which is why every title/abstract query missed them — **rule 80 in its purest form, and the previous entry's own caveat predicted this exact failure.** The measured result: **all MLL-AF4 binding occurs at open unmethylated-CpG (uCpG) regions, with the strongest uCpG enrichment at peak centres, but MLL-AF4 occupies only 20% of uCpG sites, while CFP1 occupies 50% and KDM2B occupies 89%** [E1]. **So the three readers massively co-occupy and the fusion is the most restricted of the three.** What distinguishes the fusion's subset is menin: MLL(N) and menin ChIP-seq reads at MLL-AF4 peaks correlate at **r2 = 0.96**, whereas menin versus CFP1 at CFP1 peaks is **r2 = 0.27** and KDM2B does not correlate with menin at all [E1]. PAFc is a weaker discriminator - 4,892 of PAFc sites (78%) carry no MLL-AF4 [E1]. **Read against the menin-interface hypothesis this is the strongest direct support in the graph: at genome scale, in the index fusion, in human cells, unmethylated CpG is necessary and manifestly not sufficient, and the variable that picks the fusion's address out of the uCpG pool is menin occupancy.** Two corollaries follow that the beat did not expect. First, **uCpG is not a scarce resource here** - the fusion leaves 80% of it unused - so a competition model in which readers exhaust a limited address pool is not the operative one, and the previous cycle's verdict of *not demonstrably contested* survives but for a better reason than *nobody measured*. Second, **that same abundance is what a hypermethylating co-mutation would have to eat into before it touched the fusion**, which is the quantitative hinge of `hyp-a-hypermethylating-comutation-should-antagonise-the-unmethylated-cpg-address`.
The four numbers, and what each was measured with
| quantity | value | assay | figure |
|---|---|---|---|
| MLL-AF4 binding sites that are open uCpG | all of them (heatmap); Venn undercounts a few | MLL(N) ChIP-seq x Bio-CAP-seq x ATAC-seq | 1E, 1F |
| uCpG sites bound by MLL-AF4 | 20% | same | 1F |
| uCpG sites bound by CFP1/CXXC1 | 50% | CFP1 ChIP-seq x Bio-CAP-seq | S1F |
| uCpG sites bound by KDM2B | 89% | KDM2B ChIP-seq x Bio-CAP-seq | S1F |
| MLL(N) vs menin reads at MLL-AF4 peaks | r2 = 0.96 | ChIP-seq scatter | 1I |
| menin vs CFP1 reads at CFP1 peaks | r2 = 0.27 | ChIP-seq scatter | 1J |
| PAFc sites with no MLL-AF4 | 4,892 (78%) | PAF1/LEO1 ChIP-seq | S1I |
The authors' own reading of the 20/50/89 spread is that "the CXXC domains of MLL-AF4, CFP1 and KDM2B are highly related, [so] the differences in the number of bound uCpGs may be due to other protein interactions influencing recruitment", and of the correlation that "an association with menin represents a feature that may serve to restrict MLL-AF4 recruitment to a particular subset of uCpG sites" 1. Those are the authors' words, not an inference of this entry.
Bio-CAP is itself a CXXC-based affinity reagent (Blackledge 2012), so the uCpG denominator is operationally "what a recombinant CXXC domain pulls down". That is a real limitation and it is stated in the caveat rather than hidden: the 20/50/89 split is three ChIPs against one common CXXC-defined reference, not three orthogonal definitions of unmethylated CpG.
The second measurement: the readers do not spread and the fusion does
At 149 spreading targets, MLL-AF4 and menin extend into the gene body, and the spreading domains are punctuated and demarcated by uCpG sites, most within 1-2 kb of each other, rarely beyond 4 kb (7 kb maximum observed) 1. Neither KDM2B nor CFP1 shows the spreading pattern, although both bind uCpGs. The authors' model is explicit and is a menin-interface model: promoter-bound MLL-AF4 stabilises further CXXC-mediated contacts at low-density gene-body uCpG through shared complex members menin and ENL, and "because other CXXC proteins, such as KDM2B, do not interact with complex members such as menin or ENL, promoter-bound KDM2B is not sufficient to stabilize neighboring CXXC-mediated recruitment" 1.
So there are two independent discriminations in one dataset, both pointing the same way: which uCpG sites (20% vs 50% vs 89%) and what happens at them (spreading vs not), and menin is the named variable in both.
And a methylation measurement at fusion targets in patients
Question four of this beat was "has anybody measured the methylation state of a KMT2A::AFF1 CpG-island target in the index disease?" The answer is yes, twice over, and both were reachable only from full text.
- In cells: Bio-CAP-seq in SEM directly measures non-methylated CpG at every MLL-AF4 peak 1.
- In patients: "RUNX2, MBNL1, JMJD1C, SENP6, MEF2C and ZEB2 are also hypomethylated in MLL-FP samples compared to either normal cells or other leukemias" (Figure S5B) - six named spreading targets, hypomethylated in patient material relative to two comparators 1.
The caveat that keeps this honest: Figure S5B is a supplementary panel, the per-gene statistics were not extracted here, and the patient cohort is MLL-FP rather than KMT2A::AFF1 specifically. The claim this entry makes is that the measurement exists and points toward hypomethylation at fusion targets - not that its effect size has been read off.
What this does and does not do to the previous cycle's verdict
Withdrawn: the numerator of absence-no-experiment-tests-cxxc-reader-competition-at-a-kmt2a-fusion-target. 0 of 11 becomes at least 2 of 11 (CXXC1/CFP1 and KDM2B), and the rule-of-three bound <27.3% is void because the event occurred.
Surviving, and strengthened: the conclusion that the address is not demonstrably contested. The previous entry reached it by absence; this reaches it by measurement, and the measurement is more favourable to the lab's hypothesis than the absence was. A reader who only had the absence would have believed the question was open; it is substantially answered, and answered in the project's favour.
Newly opened, and not answered by anything: co-occupancy is not competition. Kerry measures steady-state occupancy of three readers, not what happens to one when another is removed. exp-cxxc-reader-depletion-versus-kmt2a-fusion-occupancy, already in this graph, remains exactly the right experiment and is now better motivated: KDM2B is on 89% of uCpG including nearly all of the fusion's sites, so a KDM2B degron is a clean test of whether that co-residency matters.
| script | none - full-text re-read of PMID 28076791 (PMC5263239) performed by hand 2026-09-21 |
|---|---|
| status | run |
| result summary | MLL-AF4 20% / CFP1 50% / KDM2B 89% of uCpG sites in SEM; menin-MLL(N) r2 = 0.96 at fusion peaks vs menin-CFP1 r2 = 0.27 at CFP1 peaks. Corrects a 0/11 counted absence written in cycle 44 to at least 2/11. |
| system | SEM, human t(4;11) KMT2A::AFF1 B-ALL cell line - index fusion, index lineage, human |
| how it was missed | The abstract of PMID 28076791 names neither CFP1 nor KDM2B nor any occupancy percentage; it says only 'unmethylated CpG regions in the gene body'. Every title/abstract query in the cycle-44 beat was therefore correct and blind at the same time. This is rule 80 with the paper already cited 29 times in the same graph. |
inputs
[ "PMC5263239 full text, efetch db=pmc, 62,170 characters after tag-strip-then-unescape" ]
corrects
[ "absence-no-experiment-tests-cxxc-reader-competition-at-a-kmt2a-fusion-target", "hyp-the-cxxc-address-is-a-contested-resource" ]
cxxc dna kd conditions
{
"_why": "Added by king sweep 2026-09-21 (cycle 46). This entry quotes the KMT2A CXXC-DNA 44 uM. That number is a property of five choices, not of the molecule, and quoting it bare is how this graph manufactured a 1,300-fold 'discrepancy' with the menin contact.",
"value": "44 uM",
"construct": "isolated CXXC domain, KMT2A 1147-1203, untagged, 15N-labelled",
"ligand": "DNA pal oligonucleotide, one CpG element",
"method": "NMR titration",
"buffer": "50 mM phosphate, pH 7.1, 300 mM NaCl, 2 mM DTT, 25 C",
"nacl_mM": 300,
"source_pmid": "20010842",
"source_quote": "\"Determination of the dissociation constants by NMR for the CXXC domain (MLL 1147-1203) and mutants was performed by titration of 15N-labeled protein with concentrated solutions of DNA pal oligonucleotide. Experiments were carried out in 50 mM phosphate buffer (pH 7.1), 300 mM NaCl, 2 mM DTT at 25 C.\" -- Cierpicki 2010 NSMB, Methods, fetched and read from PMC2908503 on 2026-09-21. The result sentence reads \"The wild type protein yielded a Kd of 44 uM\".",
"this_is_the_high_salt_end": "300 mM NaCl is roughly twice physiological ionic strength and SIX TIMES the 50 mM the menin-MBM1 53 nM was measured in. The same domain measured tagless at 150 mM NaCl gives 5.7 uM -- 7.7-fold tighter from salt alone.",
"do_not_compare_naively": {
"against": "menin-MBM1 53 nM (PMID 20961854), measured at 50 mM NaCl by FP",
"best_matched_pair": "108-fold, not ~1,000-fold",
"see": "ana-the-menin-affinity-comparison-was-made-at-one-sixth-the-salt"
},
"added": "2026-09-21"
}- This is favourable to the menin-interface hypothesis and that is exactly why it should be attacked first. The same paper's tethering experiment shows the menin interaction is sufficient but not necessary for recruitment, so a menin inhibitor may move the fusion rather than evict it - and a fusion relocated onto the 80% of uCpG it currently ignores is an untested and unmeasured state in every menin-inhibitor trial running today.
- If KDM2B is on 89% of uCpG sites including the fusion's, then a therapeutic strategy that targets the CXXC-DNA interface pharmacologically has a selectivity problem that is quantitative and severe, not theoretical. target-phase2-cxxc-dna-dossier should carry these numbers.
- MLL-AF4 sits on 20% of uCpG sites and KDM2B on 89%. What fraction of the fusion's sites are also KDM2B-occupied, and does degrading KDM2B change fusion occupancy at them? Kerry reports the marginals and not the joint.
- If menin occupancy is what selects the fusion's 20%, does a menin inhibitor collapse the fusion onto a different subset of uCpG or off chromatin entirely? Nobody has run Bio-CAP-seq plus fusion ChIP under revumenib.
- The 33 nM SPR figure and the 44 uM NMR figure for the CXXC-DNA contact differ by ~1,300-fold. Which construct and which DNA does the fusion actually see in a cell, and does the weak-contact argument this graph makes in several entries survive the stronger number?
- Bio-CAP defines uCpG with a CXXC reagent. What do the 20/50/89 percentages look like against whole-genome bisulfite sequencing in the same cells, where the reference is independent of the domain being studied?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In SEM cells (human t(4;11) KMT2A::AFF1 B-ALL), Bio-CAP-seq plus ChIP-seq show that all MLL-AF4 binding occurs at open unmethylated-CpG regions but MLL-AF4 occupies only 20% of uCpG sites while CFP1 occupies 50% and KDM2B occupies 89%; MLL(N) and menin reads correlate at r2 = 0.96 at MLL-AF4 peaks whereas menin and CFP1 correlate at r2 = 0.27, and neither KDM2B nor CFP1 shows the gene-body spreading that MLL-AF4 and menin show.
MLL-AF4 Spreading Identifies Binding Sites that Are Distinct from Super-Enhancers and that Govern Sensitivity to DOT1L Inhibition in Leukemia. · Kerry J, Godfrey L, Repapi E, et al. · Cell Rep 2017What this evidence cannot show
Read in full text from PMC5263239 on 2026-09-21, not from the abstract, which names none of these results. One cell line, SEM, and a single set of ChIPs; no biological replicate ChIP is reported for CFP1 or KDM2B. The uCpG denominator is defined by Bio-CAP, which is itself a recombinant CXXC affinity reagent, so all three percentages are measured against a CXXC-defined reference rather than against bisulfite sequencing - the ranking (20 < 50 < 89) is more robust than the absolute values. Percentages are read from the text; underlying peak counts were not extracted and the two scatter r2 values are computed over different peak sets (MLL-AF4 peaks vs CFP1 peaks), so they are not a like-for-like comparison of correlation strength. Co-occupancy is steady state; no reader was depleted.
Evidence that argues against this (1)
- The same paper shows the menin interaction is SUFFICIENT but NOT NECESSARY for recruitment in a TetR/BAC tethering system in mouse ES cells (their Figure 2), so a correlation of r2 = 0.96 between menin and fusion occupancy cannot be read as menin being the only route to the address - and a graph built on the menin interface should hold that limit explicitly. Kerry J et al. 2017
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The KMT2A CXXC (MT) domain binds DNA containing unmethylated CpG and is blocked by 5-methyl-CpG on either strand, with an SPR KD of approximately 3.3e-8 M for a GST-MT fusion, establishing that the fusion's DNA contact is methylation-discriminating.
The MT domain of the proto-oncoprotein MLL binds to CpG-containing DNA and discriminates against methylation. · Birke M, Schreiner S, Garcia-Cuellar MP, et al. · Nucleic Acids Res 2002What this evidence cannot show
Recombinant GST fusion in vitro. GST dimerises, so the 33 nM SPR figure carries avidity and is not comparable with the ~44 uM NMR-titration figure this graph quotes elsewhere for the isolated CXXC domain on a palindromic CpG oligonucleotide - different construct, different method, different DNA, and a roughly 1,300-fold gap that this graph currently does not acknowledge in either entry. Do not reconcile them by picking one; record both with their methods. This entry uses the paper only for the qualitative methylation discrimination, which is not in dispute. RESOLVED 2026-09-21 by Q82. This caveat was right to flag the discrepancy and its GST-dimerisation attribution is now bounded rather than confirmed. Ten published measurements of this contact form a continuous series from 44 uM to 33 nM. A same-laboratory, same-construct (1147-1203), same-oligonucleotide pair puts GST plus a 300-to-150 mM salt drop at 5.5-fold, and a tagless-against-tagless pair (44 uM at 300 mM NaCl against 5.7 uM at 150 mM NaCl) puts about 8-fold on the salt alone - so GST dimerisation explains very little of the gap. The remaining 19-fold, between GST-1147-1203 at 0.64 uM and GST-1147-1244 at 33 nM, confounds the 41-residue post-CXXC basic patch, six CpGs against two, a 52-bp duplex against a 12-mer, and an immobilised kinetic geometry, and is unresolved. See ana-cxxc-dna-affinity-ladder. This entry's own use of the paper - qualitative methylation discrimination - is unaffected.
Evidence that argues against this (2)
- This graph's other affinity figure for the same interaction is ~44 uM by NMR titration on a palindromic CpG oligonucleotide, and several entries build the argument 'the CXXC contact is too weak to be the sole determinant' on it. The two numbers cannot both be the affinity of the same contact, and nothing in this graph currently reconciles them. Birke M et al. 2002
- The reconciliation this caveat called for was published in 2008, by the same laboratory that produced the 44 uM figure, and this graph did not cite it. Erfurth et al. state the SPR value is about two orders of magnitude tighter than their ITC and attribute it to GST dimerisation and bivalent binding, and the same paper gives the MT domain boundaries as MLL 1147-1244 and an ITC ladder of 3.5 to 32 uM across five Hoxa9-derived oligonucleotides at fixed construct, salt and method. Erfurth FE et al. 2008
Withdraws the 0/11 numerator: CXXC1/CFP1 and KDM2B both have published occupancy in SEM cells alongside MLL-AF4 (Kerry 2017, Figures 1 and S1F).
Supplies the missing measurement: three CXXC readers co-occupy uCpG massively (20/50/89%), so the address is shared but not scarce; the conclusion survives, the premise of absence does not.
Menin occupancy is the variable that selects the fusion's 20% out of the uCpG pool: r2 = 0.96 with MLL(N) at fusion peaks, against 0.27 for CFP1.
KDM2B occupies 89% of unmethylated CpG sites in SEM, the widest of the three readers measured, and does not spread.
CFP1/CXXC1 occupies 50% of unmethylated CpG sites in the same cells and does not correlate with menin (r2 = 0.27).
Genome-wide Bio-CAP-seq confirms the address is unmethylated CpG and quantifies how much of it the fusion actually uses.
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
Measured in SEM, a human t(4;11) KMT2A::AFF1 line - the index fusion in the index lineage, which is rare for an occupancy dataset in this graph.
Quantifies the dossier's selectivity gate: a CXXC-DNA ligand would meet KDM2B on 89% of the same sites.
Records the unreconciled affinity gap: 33 nM by SPR on a GST-MT fusion versus ~44 uM by NMR on an isolated domain.
Motivates the experiment better than the absence did: KDM2B co-resides on nearly every uCpG site the fusion uses, so a KDM2B degron is now a targeted test rather than a fishing trip.
Whether hypermethylation can reach the fusion depends on the 20% figure: the fusion leaves most unmethylated CpG unused.
The headroom reading rests entirely on the 20/50/89 measurement and the r2 = 0.96 menin correlation.
Resolves the discrepancy this entry flagged as unreconciled in its own caveat last cycle, and bounds its GST-dimerisation explanation at about 5.5-fold.
same paper, same full-text reading; that entry's figures are reconfirmed here